Personalized neuromotor rehabilitation therapy for upper limb using a neuromusculoskeletal arm model

A system comprising:

• one or more data storage devices operatively coupled to one or more hardware processors via the one or more input/output interfaces;
• and configured to store instructions configured for execution via the one or more hardware processors to:
• personalize a neuromusculoskeletal arm model having three joints and a set of muscle actuators, wherein the neuromusculoskeletal arm model is adapted to conform with anthropometric measurements of a subject and is configured to perform one or more motion tasks by the subject using the three joints being a ball and socket shoulder joint along with clavicle and scapula having three degrees of freedom (DoF), an elbow joint with one DoF and a wrist joint having three DoF;
• optimize the set of muscle actuators by performing the following iterative steps starting with the set of muscle actuators in the personalized neuromusculoskeletal arm model in a first iteration and every subsequent iteration having a reduced number of muscle actuators in the set of muscle actuators compared to a previous iteration by:
• calculating a muscle torque indicative of a muscle force based on a moment arm └R(q)┘, wherein q represent a joint angle, an activation value a obtained from a muscle Electromyography (EMG) data, a normalized length l and a normalized velocity {dot over (l)} of each muscle actuator in the set of muscle actuators;
• calculating a joint torque indicative of loading on each of the three joints, wherein the joint torque is based on an acceleration {umlaut over (q)} due to the joint torque τ, the joint angle q, an inertia tensor function Q, a mass matrix M, a Coriolis component C, a gravity component G and an external force F applied to the neuromusculoskeletal arm model;
• generating a trajectory associated with a corresponding motion profile for each of the one or more motion tasks based on forward dynamics using the calculated muscle torque and the calculated joint torque;
• and
• evaluating a mean square error between the generated trajectory and a target trajectory associated with a corresponding motion profile for each of the one or more motion tasks;
• until the evaluated mean square error is less than a pre-defined threshold;
• and
• identify the set of the muscle actuators with the reduced number of muscle actuators at the end of the iterative steps as the optimized set of muscle actuators to perform the one or more motion tasks.

A computer program product comprising a non-transitory computer readable medium having a computer readable program embodied therein, wherein the computer readable program, when executed on a computing device, causes the computing device to:

• personalize a neuromusculoskeletal arm model having three joints and a set of muscle actuators, wherein the neuromusculoskeletal arm model is adapted to conform with anthropometric measurements of a subject and is configured to perform one or more motion tasks by the subject using the three joints being a ball and socket shoulder joint along with clavicle and scapula having three degrees of freedom (DoF), an elbow joint with one DoF and a wrist joint having three DoF;
• optimize the set of muscle actuators by performing the following iterative steps starting with the set of muscle actuators in the personalized neuromusculoskeletal arm model in a first iteration and every subsequent iteration having a reduced number of muscle actuators in the set of muscle actuators compared to a previous iteration by:
• calculating a muscle torque indicative of a muscle force based on a moment arm └R(q)┘, wherein q represent a joint angle, an activation value a obtained from a muscle Electromyography (EMG) data, a normalized length l and a normalized velocity {dot over (l)} of each muscle actuator in the set of muscle actuators;
• calculating a joint torque indicative of loading on each of the three joints, wherein the joint torque is based on an acceleration {umlaut over (q)} due to the joint torque τ, the joint angle q, an inertia tensor function Q, a mass matrix M, a Coriolis component C, a gravity component G and an external force F applied to the neuromusculoskeletal arm model;
• generating a trajectory associated with a corresponding motion profile for each of the one or more motion tasks based on forward dynamics using the calculated muscle torque and the calculated joint torque;
• and
• evaluating a mean square error between the generated trajectory and a target trajectory associated with a corresponding motion profile for each of the one or more motion tasks;
• until the evaluated mean square error is less than a pre-defined threshold;
• and
• identify the set of the muscle actuators with the reduced number of muscle actuators at the end of the iterative steps as the optimized set of muscle actuators to perform the one or more motion tasks.